Does counseling increase sustained benefit of HAART among prison inmates after release to the community?

(1)

1 5 J A N U A R Y

Correspondence

Is Routine Testing Mandatory or Voluntary?

The AIDS Treatment Activists Coalition (ATAC) is a national coalition of AIDS activists, many of whom are living with HIV/AIDS, working together to end the AIDS epidemic by advancing research on HIV/AIDS. ATAC’s Access to Health Care for the Incarcerated working group has the goal of increasing access to health care for incarcerated people through strategy coordination, information sharing, train-ing, and advocacy for allied individuals, groups, and communities. The Access to Health Care for the Incarcerated working group opposes the call for routine HIV test-ing recommended in the recent editorial commentary entitled “HIV Infection Be-hind Bars” by Boutwell and Rich [1].

We believe that routine HIV testing in prison is virtually synonymous with man-datory testing. Prisoners often receive no pretest counseling or education about HIV infection; frequently, they are unaware of the implications of testing positive for HIV while incarcerated. Consent often means something completely different in prison than it does outside of prison. Many pris-oners are unaware that they have con-sented to HIV testing—or that they ac-tually were tested for HIV, despite the consent form.

We agree wholeheartedly that identifi-cation of infected individuals is an im-portant step in addressing HIV infection in any setting. However, prisoners do not have guaranteed access to the standard of care for HIV treatment, so beginning the process of addressing HIV disease behind bars must entail providing access to coun-seling, education, and the standard of care for HIV treatment, as defined by the De-partment of Health and Human Services’ guidelines [2]. Without guaranteed access to quality care, any testing will only result

in continued retaliation, stigmatization, threats, lack of confidentiality, and sub-standard medical care.

The disproportionately high number of HIV-infected persons in our prisons is more than “a challenge to correctional health systems” and a “public health op-portunity” [1]. It is a human rights ca-tastrophe of vast proportions, and it chal-lenges our society as a whole. Our priorities must be ordered by the thou-sands of prisoners whose lives are on the line.

The commentary by Boutwell and Rich [1] echoes the unfortunate new position on mandatory testing for prisoners cur-rently held by the Centers for Disease Control and Prevention (CDC). Although testing may have led to better treatment in Rhode Island (which, not surprisingly, has the smallest correctional system in the country), this is not the case in most state prison systems. The experience inside most prison systems is that testing does not lead to education, care, and treatment. It leads to discrimination, segregation, and just plain poor care. HIV-infected pris-oners are not eligible for many prison jobs and programs, such as work-release pro-grams and halfway house placement. De-spite the success of peer education in the community, behind the walls, there are fewer HIV peer education programs than ever before. The reality is that testing pos-itive inside feels like—and, in some cases, is—a death sentence.

We must ask whether it is in anyone’s interest for people inside the walls of prison to be treated any differently than people on the outside.

Mandatory HIV testing is a vast depar-ture from commonly held public health standards. Is this a precursor to a broader CDC policy that all people should be tested, regardless of consent? Right now,

people outside of prison walls have the right to make informed decisions about whether to be tested for HIV. Prisoners deserve the same rights.

Our first priority must be programming that provides education, counseling, and a continuum of quality care and treatment to inmates while they are imprisoned and after they have been released.

Acknowledgments

Potential conflicts of interest. All authors: no conflicts.

Jackie Walker,1_{Romeo Sanchez,}2,3,4 Julie Davids,5_{Mel Stevens,}6_{Laura Whitehorn,}7 Judy Greenspan,8_{and Robert Mealey}9 1_{American Civil Liberties Union National Prison}

Project, Washington, DC;2_{Alliance for Inmates with}

AIDS,3_{Coalition of Persons Living with Hepatitis C,} 4_{Latino Commission on AIDS,}5_{Community HIV/AIDS}

Mobilization Project (CHAMP),6_{The “No Time to}

Lose” Prison Committee, ACT-UP NY, and7_POZ

Magazine, New York;8_{HIV/Hepatitis C Committee}

of California Prison Focus, San Diego; and9_Student

Global AIDS Campaign, AIDS Treatment Activist Coalition, Burlington, Vermont

Reference

1. Boutwell A, Rich JD. HIV infection behind bars. Clin Infect Dis 2004; 38:1761–3. 2. The Panel on Clinical Practices for Treatment

of HIV Infection. Guidelines for the use of an-tiretroviral agents in HIV-infected adults and adolescents. U.S. Department of Health and Human Services, 2004.

Reprints or correspondence: Robert Mealey, AIDS Treatment Action Coalition Prison Health Committee, 382 Ira Pl., Bayport, NY 11705 (rmealey@smcvt.edu).

Reply to Walker et al.

Sir—We laud Mealy et al. [1] for bring-ing more attention to the complex issue of HIV testing for incarcerated popula-tions in the United States. We share their

(2)

stated goals of advancing research to end the AIDS epidemic and increasing access to health care for incarcerated people, including provision of education, coun-seling, medical care, and linkage to care in the community upon discharge for HIV-infected inmates. Our original article described the high burden of disease in the incarcerated population and recom-mended ways to address this from a med-ical and public health perspective [2]. It is true that we did not emphasize the human rights tragedy that exists in our society, in which minority and impoverished popu-lations are disproportionately incarcer-ated, and the underlying causes of racism, addiction, mental illness, and poverty are not addressed. This is a problem that we all must strive to solve.

A few key misconceptions in the letter from Mealy et al. [1] warrant specific clar-ification. The most significant misconcep-tion is their incorrect equamisconcep-tion of routine testing with mandatory testing. Testing can be voluntary (“opt in”), mandatory (no choice—all must get tested), or rou-tine (“opt out”). Each option has its pros and cons. Voluntary testing preserves maximal autonomy but fails to identify many infected individuals and fails to offer testing to noninfected individuals, a pro-cedure that, in itself, can cause individuals to reflect upon their risk behaviors and reduce them (primary prevention). Vol-untary testing also may increase stigma-tization, because only those who admit to having risk factors get tested. Mandatory testing can identify nearly all infected in-dividuals in the most timely fashion, but it does not provide the individuals with any choice. This is analogous to what hap-pens in most correctional facilities with regard to screening for tuberculosis, an-other potentially fatal, treatable contagious disease. We do not advocate mandatory HIV testing, because we believe that the same benefits (identification and treat-ment of most infected individuals, as well as some prevention) can be achieved with a well-designed system of routine testing.

Routine testing preserves the ability of the individual to opt out. There are several examples of programs that have been able to achieve testing rates of close to 90% [3], and we have found this to be well accepted by inmates. Routine HIV testing of prisoners was able to identify one-third of all known HIV-infected individuals in the entire state of Rhode Island [4].

Testing for HIV is the first step in link-age to care, whether in the community or in the correctional setting. Linkage to care includes connecting individuals with post-test counseling and education to identify opportunities to decrease personal risk of HIV acquisition or transmission and to understand the importance of accessing medical care to prolong survival.

Incarcerated individuals are among the most stigmatized and disenfranchised. As such, they are often systematically ex-cluded from accessing health services in the community. Because it is imperative to ensure linkage to care after a diagnosis of HIV infection for individuals in the community, so too is it for individuals in the correctional setting. The challenges of doing this in the two settings are unique; however, they are not insurmountable, as evidenced by model programs [5, 6]. In fact, the constitutional standard estab-lished by the US Supreme Court in Estelle v. Gamble in 1976 offers a unique oppor-tunity and requirement to provide appro-priate health care to incarcerated individ-uals, a protection that does not exist in the community setting. Thus, the en-counter with the correctional health sys-tem is an incredible opportunity to pro-vide health care to individuals from traditionally underserved and hard-to-reach communities. Some studies have found that 3 of 4 HIV-infected inmates began their first antiretroviral regimen while incarcerated [7]. With the wide-spread use of effective antiretroviral ther-apy in the late 1990s, mortality rates for HIV infection plunged in the United States. In the correctional setting, there was a parallel 75% decrease in mortality

rates [8]. Correctional health care systems are not perfect at providing care for infected persons, but clearly, many HIV-infected inmates are getting access to treatment.

Mealy and colleagues’ [1] dismissal of Rhode Island as the “smallest correctional system in the country” is neither accurate (7 states have smaller correctional popu-lations) nor on point. Model programs, such as that in Rhode Island, demonstrate feasibility and effectiveness. Large depart-ments of corrections often comprise nu-merous smaller facilities; thus, implemen-tation in other venues is more a matter of process and content than scale.

To reject use of a diagnostic test because the health care system is imperfect is to tacitly accept the shortcomings of the bro-ken system and thus to be complicit in systematically excluding individuals sub-ject to that system from accessing vital health information. This is certainly not the intention of any advocate for incar-cerated individuals. We all must insist that correctional health care services do a bet-ter job of addressing what is known to be a disproportionately high burden of in-fectious diseases among the populations they are charged to serve [9].

The time is now to address the need for improved HIV testing and linkage to care in the incarcerated population. As routine testing becomes more and more encour-aged in the community, this trend should not lag behind in the correctional setting. Indeed, this is an area where correctional health systems should be on the forefront of leadership in this epidemic. As HIV testing in high prevalence areas becomes routine in the community, there is no de-fensible reason for this service to be denied to those behind bars. The sooner that HIV-infected prisoners are aware of their diagnoses, the more likely they can begin to receive life-saving treatment and, per-haps, at the same time reduce transmission to others. Our challenge is to ensure that timely diagnosis and treatment are offered to as many individuals as possible.

(3)

Acknowledgment

Amy Boutwell, Scott Allen, and Josiah D. Rich

Brown University, The Miriam Hospital, Providence, Rhode Island

References

1. Walker J, Sanchez R, Davids J, et al. Response to the call for mandatory HIV testing in prisons [letter]. Clin Infect Dis 2004; 39:319 (in this issue).

2. Boutwell A, Rich JD. HIV infection behind bars. Clin Infect Dis 2004; 38:1761–3. 3. MacGowan RJ, Khan R, Margolis AD, et al.

HIV C&T among incarcerated young men in 5 US prisons [abstract T2-C0704]. In: Program and abstracts of the National HIV Prevention Conference (Atlanta). 2003.

4. Desai AA, Latta ET, Spaulding A, Rich JD, Flan-igan TP. The importance of routine HIV testing in the incarcerated population: the Rhode Is-land experience. AIDS Educ Prev 2002; 14(Suppl B):45–52.

5. Samet JH, Libman H, LaBelle C, et al. A model clinic for the initial evaluation and establish-ment of primary care for persons infected with human immunodeficiency virus. Arch Intern Med 1995; 155:1629–33.

6. Rich JD, Holmes L, Salas C, et al. Successful linkage of medical care and community services for HIV-positive offenders being released from prison. J Urban Health 2001; 78:279–89. 7. Altice AF, Mostashari F, Thompson AS,

Fried-land GH. Perceptions, acceptance and adher-ence to antiretrovirals among prisoners [ab-stract 253]. In: Program and ab[ab-stracts of the 4th Conference on Retroviruses and Oppor-tunistic Infections (Washington, DC). Alexan-dria, VA: Foundation for Retrovirology and Human Health, 1997.

8. Maruschak LM. HIV in prisons, 2001. Bureau of Justice Statistics Bulletin, US Department of Justice, Office of Justice Programs. NCJ 2004; 202293:1–8.

9. Hammett TM, Harmon P, Rhodes W. The bur-den of infectious disease among inmates of and releasees of US correctional facilities, 1997. Am J Pub Health 2002; 92:1789–94.

Reprints or correspondence: Dr. Josiah D. Rich, Brown Uni-versity, The Miriam Hospital, 164 Summit Ave., Providence, RI 02906.

Does Counseling Increase Sustained Benefit of HAART among Prison Inmates after Release to the Community? Sir—The lack of sustained effectiveness of HAART after release to the community of HIV-infected inmates treated in prison was well demonstrated by Springer et al. [1] in a recent article. This disappointing result occurred even though all of the pa-tients scheduled for release were referred for transitional case management services to a community-based organization and were provided with a 2-week supply of medications, a medical appointment with an HIV care provider, emergency housing and food, and assistance with other iden-tified unmet needs.

In Italy, 7.5% of the∼56,000 inmates are HIV infected (Ministry of Justice, Dipartimento dell’Amministrazione Pen-itenziaria, personal communication); it is estimated that approximately one-third of them are receiving treatment with an-tiretroviral drugs. Our data on the effec-tiveness of HAART in the prison setting are consistent with those of Springer et al. [1]; however, in some of the 220 Italian prisons, the implementation of directly observed treatment (DOT) has been found to increase the proportion of patients with undetectable HIV RNA levels [2].

We also observed that interruption of HAART after release from the prison is a common event. As suggested by Springer et al. [1] and as emphasized in the editorial commentary by Boutwell and Rich [3], there is a need for implementation of more-effective programs to reduce the probability of discontinuation of and/or reduced compliance with therapy. These programs must prevent use of illegal drugs and alcohol, homelessness, and the relapse of other behavior that may interfere with adherence to antiretroviral drug treatment.

To this end, in January 2001, we started an intensive counseling program that ad-dressed sustained correct use of HAART. During their detention, prison inmates

met with the medical staff of drug treat-ment units who were in charged with ob-serving them after release.

The study was conducted in 4 prisons on Sardinia island; the prisons were lo-cated in the towns of Alghero, Macomer, Tempio Pausania, and Sassari. Of ∼500 inmates, 16.4% were HIV infected. DOT was implemented in only 1 of the 4 pris-ons. During the 36-month study period, a total of 153 HIV-infected patients who were receiving HAART were released from prison (59 patients in 2000, 51 in 2001, and 43 in 2002). All of the subjects were injection drug users, and all attended the same drug treatment unit after prison release.

We compared the proportion of pa-tients sustaining HAART among those re-leased in the 12 months before versus the 24 months after the initiation of the coun-seling program (i.e., the year of 2000 ver-sus the years of 2001–2002). The propor-tion of patients with an HIV RNA level of !_{50 copies/mL at the time of release} from the prison increased from 20 (33.9%) of 59 subjects in 2000 to 24 (47.1%) of 51 subjects in 2001 (P1.05) and to 27 (62.8%) of 43 subjects in 2002 (OR, 3.29; 95% CI, 1.34–8.15; P!_.01). Compared with the year 2000, when only 7 (12.5%) of 56 patients were found to continue HAART, as measured by fre-quency of follow-up visits and drug dis-tribution by the community staff after the release from the prison, the proportion increased to 17 (37.8%) of 45 patients in 2001 (OR, 4.25; 95% CI, 1.43–13.01; P! ) and to 20 (52.6%) of 38 patients in .01

2002 (OR, 7.77; 95% CI, 2.56–24.66; P! ). Undectectable HIV RNA levels oc-.0001

curred in 5 (71.4%) of 7 patients in 2000, increasing to 14 (82.3%) of 17 patients in 2001 and to 16 (80%) of 20 patients in 2002.

Our data suggest that an intensive counseling program that addresses crea-tion of a relacrea-tionship between the inmate and the medical team committed to pa-tient clinical follow-up outside the prison

(4)

may improve adherence both in prison and in the community after release, at least in countries with a high proportion of in-jection drug users among HIV-infected inmates.

Acknowledgments

Sergio Babudieri,1_{Antonio Pintus,}1 Ivana Maida,1_{Giulio Starnini,}2 and Giovanni Rezza3 1_{Istituto Malattie Infettive, University of Sassari,}

and2_{Dipartimento dell’Amministrazione}

Penitenziaria, Ministry of Justice, and3_Istituto

Superiore di Sanita`, Rome, Italy

References

1. Springer SA, Pesanti E, Hodges J, Macura T, Doros G, Altice FL. Effectiveness of antiretro-viral therapy among HIV-infected prisoners: reincarceration and the lack of sustained benefit after release to the community. Clin Infect Dis

2004; 38:1754–60.

2. Babudieri S, Aceti A, D’Offizi G.P, Carbonara S, Starnini G. Directly observed therapy to treat HIV infection in prisoners. JAMA 2000; 284: 179–80.

3. Boutwell A, Rich JD. HIV infection behind bars. Clin Infect Dis 2004; 38:1761–2.

Reprints or correspondence: Dr. Sergio Babudieri, Istituto di Malattie Infettive, Universita` degli Studi di Sassari, via Giu-seppe Mazzini, 1, 07100 Sassari, Italy (babuder@uniss.it). Clinical Infectious Diseases 2005; 40:321–2 2004 by the Infectious Diseases Society of America. All rights reserved. 1058-4838/2005/4002-0023$15.00

Reply to Babudieri et al. Sir—Development of novel interventions for vulnerable populations of prisoners as they are released to the community is ur-gently needed. According to data we pre-sented, case management services alone appear to be inadequate [1]. Babudieri et al. [2] describe a novel counseling pro-gram that they believe has contributed not only to virological success in subjects in the correctional setting, but also to con-tinued success in subjects after release from prison. They should be commended for recognizing the problem and address-ing an obstacle that many view as insur-mountable. Notwithstanding the

limita-tions of their small sample size and the use of a preintervention/postintervention analytic approach, there is merit in ex-amining the lessons learned from their experience.

The high rates of viral suppression ob-served 2 years after implementation of the intervention are impressive. Several ques-tions regarding the effect of the interven-tion, however, remain. Unlike the situa-tion in the United States and elsewhere, the Sardinian experience suggests that all released prisoners are linked to continued drug treatment programs after release. The lack of a control group leads us to spec-ulate that several additional factors may have contributed to the described success. Correctional and community HIV care providers may have developed enhanced counseling skills, may have increased their diagnosis and treatment of comorbid mental illness among the released pris-oners, or may have prescribed simpler, less toxic, and more potent regimens as cli-nicians and health care systems began to embrace the importance of adherence [3]. It is also possible that sentencing laws or treatment practices changed over the ob-servation period. For instance, a greater number of prisoners released on proba-tion or parole may result in increased rates of abstinence from illicit drug use because of the stringent and often punitive con-ditions of release. A description of the types of linkages, amount of education provided, and the duration of success after release to the community would further inform us regarding the benefit of their program.

Irrespective of the extenuating circum-stances, novel approaches that sustain the benefit of antiretroviral therapy between the correctional and community settings must be developed and tested using rig-orous, unassailable, controlled trials. Only such trials will result in policy changes. We recently reported, in a randomized, con-trolled trial of directly administered anti-retroviral therapy, that supervised therapy was superior at reducing viral load and increasing CD4 lymphocyte count,

com-pared with self-administered therapy, among HIV-infected drug users [4]. The details of the intervention have been de-scribed elsewhere [5]. These data suggest that provision of structure and social sup-port provide benefit to HIV-infected drug users prescribed antiretroviral therapy— perhaps this might be applicable for re-leased prisoners. Drug treatment pro-grams that provide effective structure and support, such as prescription of metha-done or buprenorphine, are absent from most prisons and their respective release programs and should be considered in fu-ture interventions [6]. The Sardinian ex-perience appears to integrate drug treat-ment into the prison and postrelease program—a wise and likely beneficial con-tribution to success. Direct access to such structured programs, recommended by the World Health Organization [7], would likely result in improved outcomes for HIV-infected prisoners. Unfortunately, no empirical studies have been conducted to date. Interventions for released prisoners must incorporate elements that address social instability, untreated substance abuse disorders, and psychiatric condi-tions and that provide structure to oth-erwise disorganized and chaotic lives. It is incumbent on science and society to ren-der beneficial and effective antiretroviral therapy to everyone affected by HIV/ AIDS, including released prisoners.

Acknowledgments

Frederick L. Altice,1_{Sandra A. Springer,}1 and Edward Pesanti2 1_{Yale University AIDS Program, New Haven,}

and2_{University of Connecticut}

Health Sciences Center, Farmington, Connecticut

References

1. Springer SA, Pesanti E, Hodges J, Macura T, Doros G, Altice FL. Effectiveness of antiretro-viral therapy among HIV-infected prisoners: reincarceration and the lack of sustained benefit after release to the community. Clin Infect Dis

2004; 38:1754–60.

(5)

Rezza G. Does counseling increase sustained benefit of HAART among prison inmates after release to the community [letter]? Clin Infect Dis 2004; 39:321–2 (in this issue).

3. Altice FL, Friedland GH. The era of adherence to HIV therapy. Ann Intern Med 1998; 129: 503–5.

4. Altice FL, Mezger J, Bruce RD, Springer SA, Hodges J, Friedland GH. Randomized, con-trolled trial of directly administered antiretro-viral therapy (DAART) versus self-administered therapy (SAT) among HIV+ drug users [ab-stract ThPpB2093]. In: Program and ab[ab-stracts of the XVth International AIDS Conference (Bangkok). Stockholm: International AIDS So-ciety, 2004.

5. Altice FL, Mezger JA, Hodges J, et al. Devel-oping a directly administered antiretroviral therapy intervention for HIV-infected drug users: implications for program replication. Clin Infect Dis 2004; 38(Suppl 5):S376–87. 6. Smith-Rohrberg D, Bruce RD, Altice FL.

Re-search note—review of corrections-based ther-apy for opiate-dependent patients: implications for buprenorphine treatment among correc-tional populations. J Drug Issues 2004; 34: 451–80.

7. World Health Organization. Prisons, drugs and society: a consensus statement on principles, policies and practices. Berne, Switzerland: WHO (Regional Office For Europe) Health in Prisons Project and the Pompidou Group of the Council of Europe. 2001. Available at: http: //www.hipp-europe.org/downloads/england -prisonsanddrugs.pdf. Accessed 9 August 2004. Reprints or correspondence: Dr. Frederick L. Altice, Yale Uni-versity AIDS Program, 135 College St., Ste. 323, New Haven, CT 06510 (raadm@yale.edu).

Sonographic Assessment of Lipodystrophy in HIV-Infected Patients: Some Open Questions

Sir—We have read with interest the article by Asensi et al. [1] because we are con-ducting a similar study. The sonographic assessment of lipodystrophy in human im-munodeficiency virus (HIV)–infected pa-tients who are treated with HAART has been the subject of recent publications [1, 2]. Ultrasonography is a more practical method than CT scanning and dual-energy X-ray absorptiometry; its lack of ionizing radiation, low cost, availability, simplicity, and reproducibility, as well as its acceptability among patients, make it an ideal tool for diagnosis of lipodystrophy

and for early identification of subcuta-neous fat thickness reduction and visceral fat increase. We have encountered some methodological issues that need to be ad-dressed, if the potential of ultrasonogra-phy is to be fully exploited.

A good correlation between the findings of ultrasonography and CT scanning has been demonstrated for the measurement of visceral fat in patients who do not have HIV infection [3]. However, no such data are available for HIV-infected patients and with regard to subcutaneous fat thickness. Given the importance of a correlation be-tween the findings of ultrasonography and CT scanning, which remains the gold standard, ultrasonography measurements should be validated with CT scanning, at least in a preliminary group.

Subcutaneous fat thickness is related to the body mass index (BMI) in HIV-infected patients [4], who tend to have an increased amount of visceral fat when their BMI is1_{27. Although Asensi et al.} [1] claim that there is no correlation be-tween BMI, subcutaneous fat thickness, and perirenal fat diameter, all of their pa-tients had a BMI of!_{27. We believe that} the cutoff values of all measurements, par-ticularly that of perirenal fat diameter, should be readjusted for patients with a BMI of1_{27, to avoid a possible bias.}

Sex-related differences in body fat dis-tribution are still a controversial issue [5, 6]. In our cohort, crural and brachial sub-cutaneous fat thickness values were higher in women than in men, in both HIV-positive and HIV-negative patients, and we feel that this finding warrants the di-vision of the cohort into 2 subgroups. There is no such distinction made in the article by Asensi et al. [1]. In fact, the low number of female patients in their cohort could represent an additional source of bias. The reproducibility of their results in a female group is a crucial point, partic-ularly with regard to the promising pre-dictive role of perirenal fat diameter [1].

Dismetabolic abnormalities, such as dislipidemia and insulin resistance, have been described in hepatitis C virus–

infected patients [7, 8], and these abnor-malities are strictly related to the increase of visceral fat. This could be a source of bias in the evaluation of visceral fat and secondarily subcutaneous fat thickness in patients coinfected with HIV and hepatitis C virus with lipodystrophy. Because hep-atitis C virus–infected patients are nu-merous in the series of Asensi et al. [1], this bias in the assessment of perirenal fat diameter cannot be avoided if one fails to include those patients in a specific subgroup.

The ability of ultrasonography to re-place—partially or completely—other more cumbersome and expensive meth-ods needs be clearly defined, given what is at stake. The interruption of a life-saving treatment such as HAART or a switch of the regimen to nonnucleoside reverse-transcriptase inhibitors because of im-pending lipodystrophy, a serious and non-reversible condition, are very important steps to take and require the support of a proven and fully standardized method.

Acknowledgments

R. Gulizia,1_{E. Brunetti,}1_{C. Gervasoni,}2 M. Galli,2_{and C. Filice}1 1_{Division of Infectious and Tropical Diseases, IRCCS}

San Matteo, University of Pavia, Pavia, Italy; and2_{Infectious and Tropical Disease Institute,}

IRCCS Luigi Sacco, University of Milano, Milan, Italy

References

1. Asensi V, Martin-Roces E, Carton J, et al. Per-irenal fat diameter measured by echography could be an early predictor of lipodystrophy in HIV type 1–infected patients receiving highly active antiretroviral therapy. Clin Infect Dis

2004; 39:240–7.

2. Martinez E, Bianchi L, Garcia Viejo MA, et al. Sonographic assessment of regional fat in HIV-1–infected people. Lancet 2000; 356:1412–3. 3. Cucchi E, Piatti PM, Orena C, et al. Is

ech-ography an adequate method for assessing the thickness of intra-abdominal fat? A comparison with computed tomography. Radiol Med

1997; 94:329–34.

4. Muurahainen N, Pettit R, Kotler D, et al. Ab-normalities in HIV-associated lipodystrophy syndrome that vary by weight status [abstract

(6)

63]. In: Program and abstracts of the 1st In-ternational Workshop on Adverse Drug Reac-tions and Lipodystrophy in HIV (San Diego).

1999:53.

5. Taylor RW, Gold E, Manning P, Goulding A. Gender differences in body fat content are pre-sent well before puberty. Int J Obes Relat Metab Disord 1997; 21:1082–4.

6. Komiya S, Eto C, Otoki K, Teramoto K, Shi-mizu F, Shimamoto H. Gender differences in body fat of low– and high–body-mass children: relationship with body mass index. Eur J Appl Physiol 2000; 82:16–23.

7. Shintani Y, Fujie H, Miyoshi H, et al. Hepatitis C virus infection and diabetes: direct involve-ment of the virus in the developinvolve-ment of insulin resistance. Gastroenterology 2004; 126:840–8. 8. Carr A, Samaras K, Thorisdottir A, et al.

Di-agnosis, prediction, and natural course of HIV-1 protease-inhibitor associated lipodystrophy, hyperlidemia and diabetes mellitus: a cohort study. Lancet 1999; 353: 2093–9.

Reprints or correspondence: Dr. Rosario Gulizia, Div. of In-fectious and Tropical Diseases, IRCCS San Matteo University of Pavia School of Medicine, Via Taramelli 5, 27100 Pavia, Italy (sargul@katamail.com).

Reply to Gulizia et al.

Sir—We agree with Gulizia et al. [1] that the value of ultrasonography to assess the visceral fat and subcutaneous fat thickness of HIV-infected patients has to be vali-dated by comparison with other objective methods, such as CT scanning. Using a more subjective tool—the score of the body fat changes directly observed by HIV-infected patients and their physicians during every follow-up visit by use of cri-teria published elsewhere [2]—we have found a significant correlation between the score of the directly observed facial and abdominal fat changes and the echo-graphic malar (Bichat) subcutaneous fat thickness (r p 0.280 P; !_.001) and ab-dominal subcutaneous fat thickness (r p 0.180 P; !_.001). In addition, a sig-nificant correlation between the score of the directly observed abdominal fat changes and the visceral fat measured as the perirenal fat diameter (PRFD) was also found (r p 0.320 P; !_.001) (unpublished data). However, we observed no signifi-cant correlation between the score of the directly observed changes in arm or thigh

fat and the subcutaneous fat thickness of the limbs, as assessed by ultrasonography. As suggested by Gulizia et al., the body mass index has to be considered, especially for overweight patients. To avoid a pos-sible bias in using the PRFD measure, we recommend using a new adjusted PRFD, which represents the ratio PRFD:body mass index. By use of adjusted PRFD and receiver operating characteristic curves, the most discriminant pre-HAART base-line adjusted PRFD for predicting the de-velopment of lipodystrophy at the end of the 27-month follow-up period was ⭓0.118. Measurement of the adjusted PRFD had identical sensitivity but was slightly more specific than measurement of the PRFD (sensitivity, 85.3%; specificity, 77.3%) [3]. The most discriminant ad-justed PRFD for predicting the develop-ment of lipodystrophy during receipt of HAART was⭓0.182, although measure-ment of the adjusted PRFD had a discri-minant power that was identical to that of measurement of the PRFD (sensitivity, 100%; specificity, 75.7%).

We also agree with Gulizia et al. [1] that the small number of women (18 patients [24.3%]) in our series—which was mostly due to the predominance of males among intravenous drug users in Spain [4], many whom are coinfected with hepatitis C virus (29 patients [39.2%])—could represent a bias due to differences in sex, status of hepatitis C virus coinfection, and body fat distribution.

Our group and others have reported that coinfection with hepatitis C virus in-creases the level of lipodystrophy (mostly lipoatrophy) in HIV-infected patients, perhaps because of liver mitochondrial toxicity, although mitochondrial toxicity also decreases the serum levels of choles-terol and triglycerides, probably because of liver damage due to hepatitis C virus infection [5–10]. Therefore, coinfection with HIV and hepatitis C virus could have consequences for subcutaneous fat thick-ness and visceral fat changes, and this should be considered as well.

Our study was conducted to evaluate

the usefulness of several parameters, mainly echographic measurements to pre-dict lipodystrophy in a nonselected pop-ulation of HIV-infected individuals that was considered as a whole. The sample size of this pilot study was relatively small, se-verely limiting the evaluation of different subgroups that could be of interest. Fur-ther multicenter studies that include higher numbers of HIV-infected patients who are at baseline and receiving HAART, and that include specific subgroups of fe-males and people coinfected with hepatitis C virus and HIV, are needed to validate the utility of ultrasonography in the as-sessment of subcutaneous fat thickness and visceral fat in HIV-infected patients, as well as to confirm the value of PRFD and adjusted PRFD measures as predictors of lipodystrophy.

Acknowledgments

V. Asensi,1_{E. Martin-Roces,}3_{J. A. Carton,}1 J. Collazos,4_{A. Alonso,}2_{M. Medina,}2 J. M. Aburto,2_{E. Martinez,}5 and J. A. Maradona1 1_{Infectious Diseases Unit and}2_Department

of Radiology, Hospital Central de Asturias, Oviedo University School of Medicine, and3_Department

of Internal Medicine, Hospital Monte Naranco, Oviedo;4_{Infectious Diseases Unit, Hospital}

de Galda´cano, Galda´cano; and5_{Infectious Diseases}

Unit, Hospital Clinic Universitari, Barcelona, Spain

References

1. Gulizia R, Brunetti E, Gervasoni C, Galli M, and Filice C. Sonographic assessment of lip-odystrophy in HIV-infected patients: some open questions. Clin Infect Dis 2005; 40:323–4 (in this issue).

2. Lichtenstein KA, Ward DJ, Moorman AC, et al. Clinical assessment of HIV-associated lip-odystrophy in an ambulatory population. AIDS 2001; 15:1389–98.

3. Asensi V, Martin-Roces E, Carton JA, et al. Perirenal fat diameter measured by echogra-phy could be an early predictor of lipodystro-phy in HIV-infected patients receiving highly active antiretroviral therapy. Clin Infect Dis

2004; 39:240–7.

4. De la Fuente L, Bravo MJ, Barrio G, et al. Lessons from the history of the human im-mune deficiency virus/acquired imim-mune de-ficiency syndrome epidemic among Spanish

(7)

Table 1. Amino acid sequence differences in diphtheria toxin of

Corynebacterium diphtheriae, Corynebacterium ulcerans X959, C. ul-cerans A2911, and C. ulul-cerans A6361.

Amino acid position C. diphtheriae (referent) C. ulcerans X959 C. ulcerans A2911 C. ulcerans A6361

2 Ser Asn Asn Asn

14 Leu Ile Leu Leu

22 Ser Ser Leu Leu

31 Val Asp Val Val

67 Thr Ala Thr Ala

116 Val Ile Ile Ile

183 Ala Ala Glu Glu

210 Ala Ser Ser Ser

233 Val Ala Ala Ala

262 Thr Ile Thr Thr

277 Gln Gln Arg Gln

294 Thr Val Val Val

296 Pro Ser Ser Ser

305 Ala Ser Ser Ser

314 Ile Val Val Val

317 Glu Lys Lys Lys

378 Ile Leu Leu Leu

415 Leu Val Val Val

417 Asp Gly Gly Gly

421 Val Ala Ala Ala

432 Arg Lys Lys Lys

491 Gly Asp Asp Asp

492 Asp Ala Ala Ala

493 Val Thr Thr Thr

500 Ser Thr Thr Thr

518 Arg Arg Thr Arg

527 Asn Asp Asp Asp

529 Ile Thr Thr Thr

530 Ser Pro Pro Pro

531 Ser Leu Leu Leu

532 Asp Ser Ser Ser

535 Gly Asp Asp Asp

556 Phe Ala Ala Ala

558 Ile Val Val Val

NOTE. Bold face indicates a change in the amino acid class from that of the reference strain. Ala, alanine; Arg, arginine; Asn, asparagine; Asp, aspartic acid; Ile, isoleucine; Gln, glutamine; Glu, glutamic acid; Gly, glycine; Leu, leucine; Lys, lysine; Phe, phenylalanine; Pro, proline; Ser, serine; Thr, threonine; Val, valine.

drug injectors. Clin Infect Dis 2003; 37(Suppl 5):S410–5.

5. Zylberger H, Nalpas B, Pol S, et al. Is there a relationship between hepatitis C virus infec-tion and antiretroviral-associated lipoatrophy? AIDS 2000; 14:2055–65.

6. Duong M, Petit JM, Piroth L, et al. Association between insulin and hepatitis C virus chronic infection in HIV-hepatitis C virus-coinfected patients undergoing antiretroviral therapy. J Acquir Immune Defic Syndr 2001; 27:245–50. 7. Torti C, Patroni A, Tinelli C, et al. Influence of hepatitis C virus coinfection on lipid ab-normalities in HIV-positive patients after highly active antiretroviral therapy. J Acquir Immune Defic Syndr 2002; 29:315–7. 8. Collazos J, Mayo J, Ibarra S, Cazallas J.

Hy-perlipidemia in HIV-infected patients: the protective effect of hepatitis C-coinfection. AIDS 2003; 17:927–9.

9. Rodriguez-Guardado A, Maradona JA, Asensi V, et al. Hepatitis C in patients with HIV in-fection and lipodystrophy. J Acquir Immune Defic Syndr 2003; 32:348–9.

10. Collazos J, Rodriguez-Guardado A, Maradona JA, et al. Factors influencing the development of lipodystrophy in human immunodeficiency virus-infected patients. Scand J Infect Dis

2003; 35:339–40.

Reprints or correspondence: Dr. Victor Asensi, Infectious Dis-eases Unit, Hospital Central de Asturias, Oviedo University School of Medicine, Celestino Villamil s/n, 33006 Oviedo, Spain (vasensia@medynet.com).

Classical Diphtheria Caused by Corynebacterium

ulcerans in Germany: Amino Acid Sequence Differences between Diphtheria Toxins from Corynebacterium diphtheriae and

C. ulcerans

Sir—Corynebacterium ulcerans is increas-ingly recognized as an emerging pathogen in various countries, including the United States [1], the United Kingdom [2], Japan [3], The Netherlands [4], Switzerland [5], and Italy [6]. In some of these countries, C. ulcerans is now the leading cause of classical pharyngeal diphtheria [2]. The ability of C. ulcerans infection to mim-ic diphtheria is explained by the fact that C. ulcerans—similar to Corynebacte-rium diphtheriae and CorynebacteCorynebacte-rium pseudotuberculosis—carries lysogenic b-corynephages coding for the diphtheria

toxin (DT). We previously described 2 C. ulcerans isolates that were associated with cases of extrapharyngeal diphtheria (one was isolated from a cutaneous infection and the other caused lethal necrotizing si-nusitis); sequencing of the respective tox gene of both strains revealed differences in nucleotide and amino acid sequences between C. diphtheriae and C. ulcerans DTs [7].

Here, we report a case of classical diphtheria caused by C. ulcerans in a 53-year-old woman who developed severe pharyngeal swelling and a whitish pseu-domembrane. Her family practitioner obtained a swab specimen from the le-sion, which grew C. ulcerans (API code 0111326), as detected by biochemical methods (API Coryne; bioMe´rieux) and 16S rRNA gene sequencing. The presence

(8)

of DT was tested both by tox PCR and an Elek test, as described elsewhere [7]. Be-cause of the severity of the swelling, the patient was referred to an ear, nose, and throat specialist who treated the infection with 2 injections of penicillin (1.2 million units). The patient recovered completely. She had received a complete course of pri-mary immunizations during childhood; however, booster immunizations were never administered.

Because no sequence data are available for the tox gene of C. ulcerans associated with classical pharyngeal diphtheria, we sequenced the complete tox gene of this strain (X959; GenBank accession num-ber AY703827) and compared the amino acid sequence with those of a C. diphthe-riae strain (GenBank accession number V01536) and 2 C. ulcerans strains, A2911 and A6361, that cause extrapharynge-al disease (GenBank accession numbers AY141013 and AY141014, respectively) [7]. The C. diphtheriae strain and the 3 C. ulcerans strains differed from each other in 26 amino acids (homology, 95.4%), 13 of which belonged to different amino acid classes (table 1). Most differences were lo-cated in the B fragment of the DT. In 3 positions (aa14, aa31, and aa262), the DT from C. ulcerans X959 was unique, har-boring an amino acid that was not present in the other DTs. Interestingly, the DTs from both the C. diphtheriae and the C. ulcerans X959 strains have a serine in po-sition 22 of the A fragment and an alanine in position 183 of the A fragment, whereas the DTs of the C. ulcerans A2911 and A6361 strains have a leucine in position 22 of the A fragment and a glutamine po-sition 183 of the A fragment. Because these are the only differences between the DTs of the C. diphtheriae and C. ulcerans X959 strains, on the one hand, and the DTs of the 2 extrapharyngeal C. ulcerans strains, on the other hand, it might be speculated that these amino acids could be involved in pseudomembrane formation; however, loss of activity in mutated C. diphthe-riae DT has not yet been linked to aa22 or aa183.

The difference between C. diphtheriae and C. ulcerans tox genes was further cor-roborated by C. ulcerans tox–specific PCR using primers DT1 with 1467R or 1586R, respectively [7], yielding a positive PCR result only for the 3 C. ulcerans strains, and not for the C. diphtheriae strain.

To our knowledge, this is the first report of classical diphtheria caused by C. ulcer-ans in Germany since the description of this species in 1995 [8]. Moreover, the se-quence data on the isolated strain that caused classical diphtheria in the patient we describe confirm the previous results that we obtained for 2 isolates of C. ul-cerans that caused extrapharyngeal in-fections, suggesting that DTs from C. diphtheriae and C. ulcerans are different, independent from an association with pharyngeal or extrapharyngeal disease. In addition, this first sequenced DT from a C. ulcerans isolate associated with classical diphtheria indicates that C. ulcerans DT amino acid sequences are less well con-served than their C. diphtheriae counter-parts [9].

Acknowledgment

Andreas Sing,1_{Suse Bierschenk,}2 and Ju¨rgen Heesemann1 1_{National Consiliary Laboratory on Diphtheria,}

the Max von Pettenkofer-Institut fu¨r Hygiene und Medizinische Mikrobiologie, and2_{Max von}

Pettenkofer-Institut fu¨r Hygiene und Medizinische Mikrobiologie, Mu¨nchen, Germany

References

1. Anonymous. Respiratory diphtheria caused by

Corynebacterium ulcerans—Terre Haute,

Indi-ana, 1996. MMWR Morb Mortal Wkly Rep

1997; 46:330–2.

2. Efstratiou A, George RC. Laboratory guide-lines for the diagnosis of infections caused by

Corynebacterium diphtheriae and C. ulcerans.

World Health Organization. Commun Dis Public Health 1999; 2:250–7.

3. Hatanaka A, Tsunoda A, Okamoto M, et al.

Corynebacterium ulcerans diphtheria in Japan.

Emerg Infect Dis 2003; 9:752–3.

4. van Dam AP, Schippers EF, Visser LG, Peek N, Swaan CM, Kuijper EJ. Difterie door infectie met Corynebacterium ulcerans in Nederland [in

Dutch]. Ned Tijdschr Geneeskd 2003; 147: 403–6.

5. Kaufmann D, Ott P, Zbinden R. Laryngophar-yngitis by Corynebacterium ulcerans. Infection

2002; 30:168–70.

6. von Hunolstein C, Alfarone G, Scopetti F, et al. Molecular epidemiology and characteristics of

Corynebacterium diphtheriae and Corynebacte-rium ulcerans strains isolated in Italy during the

1990s. J Med Microbiol 2003; 52:181–8. 7. Sing A, Hogardt M, Bierschenk S, Heesemann

J. Detection of differences in the nucleotide and amino acid sequences of diphtheria toxin from

Corynebacterium diphtheriae and Corynebacte-rium ulcerans causing extrapharyngeal

infec-tions. J Clin Microbiol 2003; 41:4848–51. 8. Riegel P, Ruimy R, de Briel D, et al. Taxonomy

of Corynebacterium diphtheriae and related taxa, with recognition of Corynebacterium

ul-cerans sp. nov. nom. rev. FEMS Microbiol Lett

1995; 126:271–6.

9. Nakao H, Mazurova IK, Glushkevich T, Po-povic T. Analysis of heterogeneity of

Coryne-bacterium diphtheriae toxin gene, tox, and its

regulatory element, dtxR, by direct sequencing. Res Microbiol 1997; 148:45–54.

Reprints or correspondence: Dr. Andreas Sing, Max von Pet-tenkofer-Institut fu¨r Hygiene und Medizinische Mikrobiologie, Pettenkoferstraße 9a, 80336 Mu¨nchen, Germany (sing@ m3401.mpk.med.uni-muenchen.de).

Influence of GB Virus Type C and HIV Coinfection on gd T cells

Sir—GB virus type C (GBV-C) coinfec-tion has been associated with decreased mortality and better outcome for human immunodeficiency virus (HIV)–infected persons [1, 2]. However, recent studies suggest that GBV-C coinfection can be secondary to HIV-disease progression, rather than an independent prognostic factor for such progression [3]. Sev-eral immunological mechanisms have been proposed that associate GBV-C coin-fection with the interference of HIV rep-lication, such as reduction of CCR5 che-mokine receptor expression, induction of antiviral chemokines, preservation of Th1 immune-response, and other mechanisms of viral interference [4, 5]. In a recent ar-ticle by Sathar et al. [6], it has been shown that GBV-C coinfection in HIV-infected persons was associated with increased fre-quency of gd T cells. An accompanying

(9)

Figure 1. Frequencies and absolute numbers of circulating gd T cells in 50 healthy control subjects (HIV⫺) and 47 HIV-infected patients (HIV+),

of whom 34 tested negative for GB virus type C (GBV-C) RNA in plasma (GBV-C⫺) and 13 tested positive (GBV-C+). A, Percentage of circulating gd T cells. B, Absolute number of circulating gd T cells. C, Absolute number of circulating Vd1 T cells. D, Absolute number of circulating Vd2 T cells. Bars, Medians and interquartile ranges. Comparison between GBV-C–infected and GBV-C–uninfected HIV patients groups was made by a nonparametric Mann-Whitney test (Prism 4; GraphPad Software). NS, P not significant.

editorial commentary [7] suggested that GBV-C coinfection may trigger a gd T cell immune response that, in turn, may in-duce a slower HIV replication rate and/or a more effective anti-HIV Th1 immune response.

Since we previously observed that a subset of gd T cells expressing the Vd2 T cell receptor may play a protective role during HIV infection by producing non-cytolytic antiviral factors [8], and that this gd T cell subset is generally depleted in HIV-infected patients [9], we have in-vestigated whether GBV-C coinfection was associated with altered frequencies of gd T cell subsets in a cohort of HIV-infected patients, which may account for the beneficial influence of GBV-C coin-fection on the course of HIV disease. The frequency of circulating gd T cells and their subset composition was analyzed by flow cytometry in 47 HIV-infected pa-tients (28 male and 19 female; median

age, 41 years; age range, 28–66 years) at-tending the outpatient facility at the Na-tional Institute for Infectious Diseases “Lazzaro Spallanzani” in Rome. In our cohort, 41 subjects (87.2% of 47) had been receiving successful antiretroviral treatment for at least 2 years prior to the study. The presence of GBV-C RNA in plasma was determined by nested RT-PCR, according to a previously estab-lished method [1]. gd T cell distribution in such subjects was compared according to their GBV-C infection status (figure 1). The 2 study groups were similar with respect to their demographic character-istics (data not shown). To establish ref-erence values for gd T cell distribution in peripheral blood, 50 blood samples from healthy donors were obtained from the Immunohaematology Laboratory at the Forlanini-San Camillo Hospital in Rome.

As shown in figure 1, the gd T cell

sub-set distribution was significantly altered by HIV infection, as expected (figure 1, pan-els C–D). In fact, the Vd1 T cell subset was significantly augmented in the pe-ripheral blood of HIV-infected subjects [9, 10]. Furthermore, the increase of Vd1 T cells was concomitant with a decrease of Vd2 T cells [9, 11, 12]. However, com-parison of gd T cell subset distribution in HIV-infected subjects, grouped according to their GBV infection status indicated that the increase in gd T cells occurred independently of GBV-C coinfection (fig-ure 1, panel C) and that GBV-C coinfec-tion had no role in the specific delecoinfec-tion of T cell receptors (figure 1, panel D).

In our cohort of HIV-infected subjects, we observed that Vd2 T cell exhaustion and Vd1 T cell increase are not correlated with GBV-C coinfection. Thus, other mechanisms are likely to be involved in interactions between GBV-C and HIV infections.

(10)

Table 1. Comparison of mean lymphocyte subset indices in HIV-infected patients with and without GB virus type C infection.

Lymphocyte subset Mean value SD P Infected subjects (n p 22) Uninfected subjects (n p 33) CD4 Percentage 26.70 8.30 26.64 8.67 NS Count, cells/mL 461.12 163.28 478.42 181.22 NS CD8 Percentage 48.47 10.08 48.70 10.78 NS Count, cells/mL 680.83_320.36 862.52_354.48 NS CD4:CD8 ratio 0.58_0.25 0.59_0.28 NS CD3 Percentage 80.0_4.17 70.99_19.76 .015 Count, cells/mL 1395.45 366.79 1451.29 571.37 NS CD30 Percentage 35.45 17.86 50.59 9.20 .041 Count, cells/mL 577.75 307.64 823.88 392.00 NS gd T Percentage 3.22 1.30 2.15 1.08 .052 Count, cells/mL 53.11 18.35 39.15 29.12 NS NOTE. NS, not significant.

Acknowledgment

Federico Martini,1_{Mariacarmela Solmone,}2 Chiara Agrati,1_{Maria R. Capobianchi,}2 Fabio Iacomi,3_{Giorgio Antonucci,}3 and Fabrizio Poccia1 1_{Cellular Immunology Laboratory,} 2_{Virology Laboratory, and}3_{1st Clinical Care Division,}

National Institute for Infectious Diseases “Lazzaro Spallanzani” IRCCS, Rome, Italy

References

1. Tillmann HL, Heiken H, Knapik-Botor A, et al. Infection with GB virus C and reduced mortality among HIV-infected patients. N Engl J Med 2001; 345:715–24.

2. Williams CF, Klinzman D, Yamashita TE, et al. Persistent GB virus C infection and survival in HIV-infected men. N Engl J Med 2004; 350: 981–90.

3. Bjorkman P, Flamholc L, Naucler A, Molne-gren V, Wallmark E, Widell A. GB virus C during the natural course of HIV-1 infection: viremia at diagnosis does not predict mortal-ity. AIDS 2004; 18:877–86.

4. Nunnari G, Nigro L, Palermo F, et al. Slower progression of HIV-1 infection in persons with GB virus C coinfection correlates with an in-tact T-helper 1 cytokine profile. Ann Intern Med 2003; 139:26–30.

5. Xiang J, George SL, Wunschmann S, Chang Q, Klinzman D, Stapleton JT. Inhibition of HIV-1 replication by GB virus C infection through increases in RANTES, 1a, MIP-1b, and SDF-1. Lancet 2004; 363:2040–6. 6. Sathar MA, York DF, Gouws E, Coutsoudis A,

Coovadia HM. GB virus type C coinfection in HIV-infected African mothers and their in-fants, KwaZulu Natal, South Africa. Clin In-fect Dis 2004; 38:405–9.

7. Bollinger RC, Gupta A. GB virus type C: a virus in search of a disease or a role in HIV therapy? Clin Infect Dis 2004; 38:410–1. 8. Poccia F, Battistini L, Cipriani B, et al.

Phos-phoantigen-reactive Vg9Vd2 T lymphocytes suppress in vitro human immunodeficiency virus type 1 replication by cell-released an-tiviral factors including CC chemokines. J Infect Dis 1999; 180:858–61.

9. Martini F, Urso R, Gioia C, et al. Gammadelta T-cell anergy in human immunodeficiency vi-rus–infected persons with opportunistic in-fections and recovery after highly active an-tiretroviral therapy. Immunology 2000; 100: 481–6.

10. Agrati C, D’Offizi G, Narciso P, et al. Gam-madelta T cell activation by chronic HIV in-fection may contribute to intrahepatic vdelta1 compartmentalization and hepatitis C virus disease progression independent of highly ac-tive antiretroviral therapy. AIDS Res Hum Re-troviruses 2001; 17:1357–63.

11. Poccia F, Boullier S, Lecoeur H, et al. Periph-eral V gamma 9/V delta 2 T cell deletion and anergy to nonpeptidic mycobacterial antigens in asymptomatic HIV-1-infected persons. J Immunol 1996; 157:449–61.

12. Poles MA, Barsoum S, Yu W, et al. Human immunodeficiency virus type 1 induces per-sistent changes in mucosal and blood gam-madelta T cells despite suppressive therapy. J Virol 2003; 77:10456–67.

Reprints or correspondence: Dr. Federico Martini, Unit of Cel-lular Immunology and Flow Cytometry, Padiglione Del Vecchio, National Institute for Infectious Diseases “Lazzaro Spallan-zani,” Via Portuense 292, 00149 Rome, Italy (martini@inmi.it). Clinical Infectious Diseases 2005; 40:326–8 2004 by the Infectious Diseases Society of America. All rights reserved. 1058-4838/2005/4002-0028$15.00

Reply to Martini et al.

Sir—We read with interest the findings of Martini et al. [1] regarding the influence of GB virus type C (GBV-C) coinfection of gd T cells in human immunodeficiency virus (HIV)–infected patients. The use of direct human-to-human microbial trans-mission as therapy for infectious diseases may seem anachronistic, but several stud-ies have suggested that infection with 1 microbe can alter the disease expression of another pathogen, including HIV [2]. To date, 11 clinical studies have associated

GBV-C coinfection with slower progres-sion to acquired immunodeficiency syn-drome, longer life, and improved response to antiretroviral therapy in HIV-infected individuals [3–6]. Several mechanisms have been proposed: namely, the induc-tion of anti-HIV cytokines and chemo-kines [3, 4, 7], the stimulation of anti-HIV cytotoxic T lymphocyte response [8], and the preservation of Th1 lymphocyte re-sponse [9]. In our limited observational study [10], we reported no significant dif-ferences in lymphocyte subsets between HIV-infected African mothers coinfected with GBV-C and those not coinfected (ta-ble 1). We observed no significant differ-ence in absolute gd T cell counts between HIV-infected mothers coinfected with GBV-C and those not coinfected. How-ever, we reported a marginally significant difference (P!_.052) between the relative gd T cell counts of the 2 groups (table 1). On the basis of this observation and the cognizance of the potential role of gd T cells in HIV infection [11], we hypothe-sized that another anti-HIV mechanism may be associated with GBV-C infection. Martini et al. [1], on the other hand,

(11)

re-port a marginally significant difference (P!_.05) in absolute gd T cell counts [1, figure 1B] and no significant difference in relative counts [1, table 1A] between the 2 groups. Although Martin et al. [1] re-ports a significant difference in gd T cell subsets between the study groups, we did not measure gd T cell subsets. In addition, there are other differences between our respective studies. In their study [1], 41 (87.2%) of 47 HIV-infected subjects had been receiving successful antiretroviral treatment for a period of 2 years. In our study [10], all HIV-infected African moth-ers were antiretroviral-naive; the only cri-terion that determined whether patients could participate was HIV infection status; the study populations are composed of different racial groups; and the majority of HIV and GBV-C infections in KwaZulu Natal are clade C and genotype 5, respec-tively. Our data concurs with the obser-vations reported by Martini et al. [1], and we agree that some other factor or factors besides GBV-C infection may play a role in altering the gd T lymphocyte subsets in HIV-infected individuals.

Acknowledgment

Mahomed A. Sathar1 and Hoosen M. Coovadia2

Departments of1_{Medicine and}2_Paediatrics

and Child Health, Nelson R. Mandela School of Medicine, Faculty of Health Sciences, University of KwaZulu-Natal, Durban, South Africa

References

1. Martini F, Solmone M, Agrati C, et al. Influ-ence of GBV-C and HIV coinfection on gd T cells [letter]. Clin Infect Dis 2004; 40:326–8 (in this issue).

2. Aronoff DM. Using live pathogens to treat infectious diseases: a historical perspective on the relationship between GB virus C and HIV. Antivir Ther 2002; 7:73–80.

3. George SL, Wunschmann S, McCoy J, Xiang J, Stapleton JT. Interactions between GB virus type C and HIV. Curr Infect Dis Rep 2003; 4: 550–8.

4. Stapleton JT. GB virus type C/hepatitis G vi-rus. Semin Liver Dis 2003; 23:137–48. 5. Williams CF, Klinzman D, Yamashita TE, et

al. Persistent GB virus C infection and survival in HIV-infected men. N Engl J Med 2004; 350: 981–90.

6. Rodriguez B, Woolley I, Lederman MM, Zdu-nek D, Hess G, Valdez H. Effect of GB virus C coinfection on response to antiretroviral treatment in human immunodeficiency vi-rus-infected patients. J Infect Dis 2003; 187: 504–7.

7. Xiang J, George SL, Wunschmann S, et al. Inhibition of HIV-1 replication by GB virus C infection through increases in RANTES, MIP-1a, MIP-1b, and SDF-1. Lancet 2004; 363:2040–6.

8. Mosier DE, Chisari FV. GB virus C mortality from HIV infection. N Engl J Med 2002; 346: 377–9.

9. Nunnari G, Nigro L, Palmero F, et al. Slower progression of HIV-1 infection in persons with GB Virus C coinfection correlates with an in-tact T-helper 1 cytokine profile. Ann Intern Med 2003; 139:26–30.

10. Sathar MA, York DF, Gouws E, et al. GB virus type C coinfetion in HIV–infected African mothers and their infants, KwaZulu Natal, South Africa. Clin Infect Dis 2004; 38:405–9. 11. Bollinger RC, Gupta A. GB virus type C: a virus in search of a disease or a role in HIV therapy? Clin Infect Dis 2004; 38:410–11. Reprints or correspondence: Dr. Mahomed A. Sathar, Dept. of Medicine, Nelson R. Mandela School of Medicine, Uni-versity of KwaZulu-Natal, PO Box 7, Congella, Durban, South Africa 4013 (sathar@nu.ac.za).

Coinfection with HIV and Human T Lymphotropic Virus Type 1: What Is the Real Impact on HIV Disease? Sir—We read with interest the article by Beilke et al. [1] summarizing their find-ings of an observational study of a group of HIV-infected patients, some of whom were coinfected with human T lympho-tropic virus types 1 (HTLV-1) and/or 2 (HTLV-2). The authors focus the clin-ical outcomes and survival probabilities among coinfected patients, and the final conclusions reinforce previous evidence of the absence of a significant impact of HTLV-2 coinfection on AIDS progression [2, 3]. They conclude that HTLV-2 co-infection is associated with some degree of protection against disease progression for such patients, with a consequent greater duration of survival. Other find-ings include the increased incidence of

some clinical events among coinfected pa-tients, such as myelopathy, urinary tract infection, peripheral neuropathy, throm-bocytopenia, and bronchitis, compared with persons who were infected with HIV alone. In addition, Beilke et al. [1] confirm already-reported data on higher CD4+_cell

counts among coinfected patients [4]. However, the absence of an impact of HTLV-1 coinfection on survival was an unexpected finding. Our previous expe-rience in evaluating a similar population with a similar duration of follow-up yielded the opposite result. We evaluated the duration of survival for 132 HIV-infected patients (63 of whom where co-infected with HTLV-1) who were observed at our clinics, with a significant difference in the time to death that favored HIV-monoinfected subjects, which remained after adjustment for injection drug use [5]. One possible explanation for such dis-crepancies could reside in the population studied: our study included mainly pa-tients in the pre-HAART era. The use of antiretroviral drugs probably modifies the natural history of coinfection, because the proper therapy for HIV infection could mask the potentially additive effects of HTLV-1 coinfection on disease progres-sion. Beilke et al. [1] stated that they ad-justed the analysis for the use of antiret-roviral drugs, but there is no description in the article on the duration of therapy or follow-up or on the time that therapy was started for monoinfected and co-infected patients. In addition, the strati-fication of CD4+_{cell counts into 3 groups}

(!_{200, 200–500, and} 1_{500 cells/mL) can} mask some important difference between HIV-monoinfected and HIV-HTLV–co-infected patients. The range of CD4+_cell

counts in the 3 groups is large enough to include some quite different patients.

Another important question regards the matching of patients by CD4+_{cell count:}

if Beilke et al. [1] conclude that this pa-rameter is modified by coinfection, then we do not think it is the best approach to compare groups. Because CD4+

cell counts are expected to be higher in HIV-HTLV–

(12)

coinfected patients, and because coinfec-tion can be associated with immune dys-function, matching patients by immune markers is not a good strategy. The au-thors recognize that they were not able to define the best threshold point to rec-ommend the commencement of HIV therapy for such patients for these reasons. The study is very important because it raises this and other questions. The sta-tistical analysis is quite strong and well designed, but some of the study’s points remained unclear. Although the authors recognize some limitations of the study, they affirm the absence of a clinical impact of coinfection on AIDS progression. We believe that the available data are still too weak to confirm that theory. As correctly stated at the end of the article, these ques-tions probably will be answered only through prospective studies.

Acknowledgments

Carlos Brites, Adriano S. Oliveira, and Eduardo M. Netto

Federal University, Bahia, Brazil

References

1. Beilke MA, Theall K, O’Brien M, et al. Clinical outcomes and disease progression among pa-tients coinfected with HIV and human T lym-photropic virus types 1 and 2. Clin Infect Dis

2004; 39:256–63.

2. Hershow RC, Galai N, Fukuda K, et al. An international collaborative study of the effects of coinfection with human T-lymphotropic vi-rus type II on human immunodeficiency vivi-rus type 1 disease progression in injection drug users. J Infect Dis 1996; 174:309–17. 3. Giacomo M, Franco EG, Claudio C, et al.

Hu-man T-cell leukemia virus type II infection among high groups and its influence on HIV-1 disease progression. Eur J Epidemiol HIV-1995; 11:527–33.

4. Schechter M, Harrison LH, Halsey NA, et al. Coinfection with human T-cell lymphotropic virus type I and HIV in Brazil—impact on markers of HIV disease progression. JAMA

1994; 271:353–7.

5. Brites C, Alencar R, Gusmao R, et al. Co-in-fection with HTLV-1 is associated with a shorter survival time for HIV-1–infected patients in Ba-hia, Brazil. AIDS 2001; 15:2053–5.

Reprints or correspondence: Dr. Carlos Brites, Universidade Federal da Bahia, Laborato´rio de Retrovirus, Rua Joa˜o das Botas, SN, Canela, Salvador, Bahia, CEP-40110-160, Brazil (crbrites@ufba.br).

Reply to Brites et al.

Sir—We appreciate the correspondence by Brites et al. [1] and would like to pro-vide additional clarification concerning several points they raised. Our study [2], which focuses on clinical outcomes and survival probabilities among patients coinfected with HIV and human T lym-photropic virus types 1 and 2 (HTLV-1 and HTLV-2, respectively), actually ar-rives at conclusions that are somewhat different than those of the previous stud-ies cited by Brites et al. [1]. In contrast to the reports of Hershow et al. [3] and Giacomo et al. [4], who found no dif-ferences in survival outcomes, our find-ings indicate improved duration of sur-vival and delayed progression to AIDS among patients coinfected with HIV and HTLV-2. A similar trend was observed among patients coinfected with HIV and HTLV-1, but the results did not reach statistical significance. Several reasons may exist for observed differences among various cohort studies. First, treatment effects at different study sites could have an impact on survival trends and rates of disease progression. In the New Orleans cohort analysis, we were careful to gen-erate a matched cohort study, in which exposure to antiretroviral therapy and CD4+ _{cell count was controlled for at}

baseline, as well as over time. The mean durations (SD) of exposure to antiret-roviral therapy among HIV–HTLV-1– coinfected patients, HIV–HTLV-2–coin-fected patients, and HIV-monoinHIV–HTLV-2–coin-fected patients were 20.02 30.10 24.21 , , and months, re-37.03 22.79 34.23

spectively. The year of initiation of anti-retroviral therapy did not differ signifi-cantly among the 3 groups.

Second, the prior studies did not report whether there was control for baseline

CD4+

cell count or viral load. We found no differences in baseline plasma HIV RNA levels (although viral load data were only available for patients who entered the cohort after 1996), and our case-con-trol cohort was matched for baseline CD4+_{cell count. We elected to stratify by}

CD4+

cell count category (i.e., !_200, 200–500, and1_{500 cells/mL). Brites et al.} [1] expressed concern that these ranges might be too wide. Therefore, we ex-amined the mean and median CD4+_cell

count range among the 3 strata, and we found almost identical results among the groups. In the CD4+ _{cell count stratum}

of 200–500 cells/mL, for example, the mean (SD) CD4+ _{cell counts were}

cells/mL (median, 362.0 345.3 77.1

cells/mL; range, 212–497 cells/mL) for HIV–HTLV-1–coinfected patients, 344.7 84.8 cells/mL (median, 340.0 cells/mL; range, 200–500 cells/mL) for HIV–HTLV-2–coinfected patients, and 350.0 83.7 cells/mL (median, 340.1 cells/mL; range, 202–494 cells/mL) for HIV-monoinfected patients. Although coinfected and mon-oinfected patients were matched with re-spect to CD4+_{cell count at baseline, the}

CD4+_{cell count could have changed over}

time (which we also demonstrated), and, therefore, the CD4+_{cell count over time}

was also controlled for in the survival analyses. We contend that matching for baseline CD4+

cell count is a rational method to compare survival and disease progression, because there may be qual-itative and quantqual-itative differences in CD4+_{cell count that account for}

diver-gences in outcomes.

In our analysis of mortality and dura-tion of AIDS-free survival, we attempted to account for potential survivor-treat-ment selection bias by matching for CD4+_{cell count at baseline/time of clinic}

entry. This is one approach that is com-monly used in observational studies [5] to eliminate potential survival bias, be-cause the final model compares persons in the different treatment or exposure groups who started in similar physical conditions and (theoretically) survive to